Supermodularity in Two-Stage Distributionally Robust Optimization --- Theory and Applications

Topic:

Supermodularity in Two-Stage Distributionally Robust Optimization --- Theory and Applications

Time&Date: 

 10:30 am -12:00 pm, 2019/12/9 (Monday)

Venue:

 Room 619, Teaching A

Speaker:

 Dr. LONG Zhuoyu (The Chinese University of Hong Kong)

Abstract:

Driven by several classic operations management problems (e.g., appointment scheduling), we solve a class of two-stage distributionally robust optimization problems which have the property of supermodularity.  We exploit the explicit upper bounds on the expectation of supermodular functions and derive the worst-case distribution for the robust counterpart. This enables us to develop an efficient method to derive an exact optimal solution of these two-stage problems. Further, we provide a necessary and sufficient condition to check whether any given two-stage optimization problem has supermodularity.  We apply this framework to classic problems, including the multi-item newsvendor problem, the appointment scheduling problem and general assemble-to-order (ATO) systems.  While these problems are typically computationally challenging, they can be solved efficiently using our approach.